Ulanqab is becoming a major source of computing power for China’s AI industry by pairing abundant renewable energy with a cool climate. Galaxy’s central data centre building covers 120,000 square metres, roughly 20 standard football pitches, and is designed to connect renewable generation, storage, power management...
Published byEdited with GPT-5.6 LunaImages generated with GPT Image 1.5
Research answer

Create a landscape editorial hero image for this Studio Global article: How is Ulanqab, a city several hundred kilometres northwest of Beijing in Inner Mongolia whose Mongolian name means “red cliff pass,” transf. Article summary: Ulanqab is converting its renewable-energy advantage into AI infrastructure: its cool climate lowers data-centre cooling needs, while abundant, relatively cheap wind and solar power support electricity-intensive computin. Topic tags: general, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fa
Ulanqab is becoming an important part of China’s AI infrastructure story. The city in Inner Mongolia has attracted data-centre investment because it combines abundant wind power, relatively low electricity costs and a cool climate—advantages that matter when computing facilities need large, continuous supplies of electricity. 20
21
22
The most visible symbol of that change is Envision Group’s Galaxy Campus, which entered operation on August 6, 2026. The project links renewable-energy generation with high-density AI computing, shifting Ulanqab’s role from a remote energy-producing region toward a potential centre for AI training and inference. 4
21
AI data centres are electricity-intensive, making energy availability and cost central to where they are built. Ulanqab’s renewable resources give operators access to substantial wind and solar generation, while its cooler conditions can make the surrounding environment more suitable for data-centre operations. 20
21
22
The location also fits a broader effort to move computing capacity toward China’s inland energy regions. Instead of sending all locally generated electricity to distant demand centres, projects such as Galaxy aim to use some of that power near its source for AI workloads. That makes energy infrastructure and computing infrastructure part of the same investment strategy.
Ulanqab’s growth is not limited to one campus. China Daily reported that the city had attracted 89 data-centre projects, while separate reporting found that electricity consumption from internet data services—primarily large data centres—rose 85.2% year over year to 4 billion kilowatt-hours in the first half of 2026. 19
23
Galaxy’s main AI data-centre building spans 120,000 square metres, or approximately the area of 20 standard football pitches. Envision describes it as the world’s largest single AI data-centre building, a characterization that should be treated as the company’s claim because rankings depend on the definition of “largest.” 1
4
6
The campus is designed to combine:
Envision says the completed design can accommodate up to 1 million AI accelerators and scale beyond 2 GW of power capacity. Those figures describe the project’s intended ceiling, not the equipment and power already operating. 1
4
10
The initial commissioned phase is reported at 120 megawatts and had been allocated to two local technology companies. In other words, Galaxy is operational, but its first live phase is considerably smaller than the campus’s eventual ambition. 5
21
The significance of Galaxy is not simply its floor area. Its proposed architecture puts power sourcing at the centre of the computing design. Envision says the campus is powered directly by renewable energy and integrates generation, storage, power management and computing in one system. 4
6
That model could help address one of the defining constraints of AI expansion: building enough power capacity quickly enough to support increasingly dense clusters of accelerators. It also gives an energy company such as Envision a direct role in the AI infrastructure market, rather than limiting it to supplying electricity or equipment.
The environmental implications require caution. Renewable-powered computing can reduce reliance on fossil-fuel generation, but the sources provided here do not establish Galaxy’s full lifecycle emissions, renewable-energy mix, utilisation rate or backup-power requirements. The strongest supported conclusion is narrower: Envision is attempting to make renewable generation a built-in part of the campus’s operating model. 3
4
Galaxy’s planned capacity is often compared with xAI’s Colossus 2 facility in the United States. Envision’s project is designed to exceed 2 GW, while Epoch AI estimates Colossus 2 at 946 MW of current IT power and projects 1,531 MW in the first quarter of 2027. 5
26
On those headline figures, Galaxy’s planned power capacity would be about one-third higher than Epoch AI’s projected Colossus 2 IT-power figure. But this is not a like-for-like measurement:
The comparison is therefore best understood as an indication of scale, not a definitive declaration that one site is already more powerful than the other.
Ulanqab’s shift illustrates how AI is changing the geography of infrastructure. The winning locations may not be the biggest coastal technology centres; they may be places with available land, reliable electricity, renewable resources and temperatures that help manage data-centre operations.
For China, the Galaxy Campus represents an attempt to connect interior energy resources with the rapidly growing demand for AI computing. For the global industry, it is another sign that the next phase of AI competition will be shaped as much by power generation, grid access and cooling as by algorithms and chips.
The project’s immediate reality is more measured than its headline targets: a 120-MW phase is operating, while the 1-million-accelerator and more-than-2-GW figures remain build-out goals. Even with that distinction, Galaxy makes Ulanqab a notable test of whether renewable-energy regions can become strategic computing hubs rather than simply exporters of power.
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
Ulanqab is becoming a major source of computing power for China’s AI industry by pairing abundant renewable energy with a cool climate.
Ulanqab is becoming a major source of computing power for China’s AI industry by pairing abundant renewable energy with a cool climate. Galaxy’s central data centre building covers 120,000 square metres, roughly 20 standard football pitches, and is designed to connect renewable generation, storage, power management and high density computing in one ca...
At its planned power capacity, Galaxy would be larger than Epoch AI’s projected 1.531 GW IT power figure for xAI’s Colossus 2 in Q1 2027, although the comparison is imperfect because the projects use different measure...
Ulanqab is becoming a major source of computing power for China’s AI industry by pairing abundant renewable energy with a cool climate. Galaxy’s central data centre building covers 120,000 square metres, roughly 20 standard football pitches, and is designed to connect renewable generation, storage, power management...
Published byEdited with GPT-5.6 LunaImages generated with GPT Image 1.5
Research answer

Create a landscape editorial hero image for this Studio Global article: How is Ulanqab, a city several hundred kilometres northwest of Beijing in Inner Mongolia whose Mongolian name means “red cliff pass,” transf. Article summary: Ulanqab is converting its renewable-energy advantage into AI infrastructure: its cool climate lowers data-centre cooling needs, while abundant, relatively cheap wind and solar power support electricity-intensive computin. Topic tags: general, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fa
Ulanqab is becoming an important part of China’s AI infrastructure story. The city in Inner Mongolia has attracted data-centre investment because it combines abundant wind power, relatively low electricity costs and a cool climate—advantages that matter when computing facilities need large, continuous supplies of electricity. 20
21
22
The most visible symbol of that change is Envision Group’s Galaxy Campus, which entered operation on August 6, 2026. The project links renewable-energy generation with high-density AI computing, shifting Ulanqab’s role from a remote energy-producing region toward a potential centre for AI training and inference. 4
21
AI data centres are electricity-intensive, making energy availability and cost central to where they are built. Ulanqab’s renewable resources give operators access to substantial wind and solar generation, while its cooler conditions can make the surrounding environment more suitable for data-centre operations. 20
21
22
The location also fits a broader effort to move computing capacity toward China’s inland energy regions. Instead of sending all locally generated electricity to distant demand centres, projects such as Galaxy aim to use some of that power near its source for AI workloads. That makes energy infrastructure and computing infrastructure part of the same investment strategy.
Ulanqab’s growth is not limited to one campus. China Daily reported that the city had attracted 89 data-centre projects, while separate reporting found that electricity consumption from internet data services—primarily large data centres—rose 85.2% year over year to 4 billion kilowatt-hours in the first half of 2026. 19
23
Galaxy’s main AI data-centre building spans 120,000 square metres, or approximately the area of 20 standard football pitches. Envision describes it as the world’s largest single AI data-centre building, a characterization that should be treated as the company’s claim because rankings depend on the definition of “largest.” 1
4
6
The campus is designed to combine:
Envision says the completed design can accommodate up to 1 million AI accelerators and scale beyond 2 GW of power capacity. Those figures describe the project’s intended ceiling, not the equipment and power already operating. 1
4
10
The initial commissioned phase is reported at 120 megawatts and had been allocated to two local technology companies. In other words, Galaxy is operational, but its first live phase is considerably smaller than the campus’s eventual ambition. 5
21
The significance of Galaxy is not simply its floor area. Its proposed architecture puts power sourcing at the centre of the computing design. Envision says the campus is powered directly by renewable energy and integrates generation, storage, power management and computing in one system. 4
6
That model could help address one of the defining constraints of AI expansion: building enough power capacity quickly enough to support increasingly dense clusters of accelerators. It also gives an energy company such as Envision a direct role in the AI infrastructure market, rather than limiting it to supplying electricity or equipment.
The environmental implications require caution. Renewable-powered computing can reduce reliance on fossil-fuel generation, but the sources provided here do not establish Galaxy’s full lifecycle emissions, renewable-energy mix, utilisation rate or backup-power requirements. The strongest supported conclusion is narrower: Envision is attempting to make renewable generation a built-in part of the campus’s operating model. 3
4
Galaxy’s planned capacity is often compared with xAI’s Colossus 2 facility in the United States. Envision’s project is designed to exceed 2 GW, while Epoch AI estimates Colossus 2 at 946 MW of current IT power and projects 1,531 MW in the first quarter of 2027. 5
26
On those headline figures, Galaxy’s planned power capacity would be about one-third higher than Epoch AI’s projected Colossus 2 IT-power figure. But this is not a like-for-like measurement:
The comparison is therefore best understood as an indication of scale, not a definitive declaration that one site is already more powerful than the other.
Ulanqab’s shift illustrates how AI is changing the geography of infrastructure. The winning locations may not be the biggest coastal technology centres; they may be places with available land, reliable electricity, renewable resources and temperatures that help manage data-centre operations.
For China, the Galaxy Campus represents an attempt to connect interior energy resources with the rapidly growing demand for AI computing. For the global industry, it is another sign that the next phase of AI competition will be shaped as much by power generation, grid access and cooling as by algorithms and chips.
The project’s immediate reality is more measured than its headline targets: a 120-MW phase is operating, while the 1-million-accelerator and more-than-2-GW figures remain build-out goals. Even with that distinction, Galaxy makes Ulanqab a notable test of whether renewable-energy regions can become strategic computing hubs rather than simply exporters of power.
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
Ulanqab is becoming a major source of computing power for China’s AI industry by pairing abundant renewable energy with a cool climate.
Ulanqab is becoming a major source of computing power for China’s AI industry by pairing abundant renewable energy with a cool climate. Galaxy’s central data centre building covers 120,000 square metres, roughly 20 standard football pitches, and is designed to connect renewable generation, storage, power management and high density computing in one ca...
At its planned power capacity, Galaxy would be larger than Epoch AI’s projected 1.531 GW IT power figure for xAI’s Colossus 2 in Q1 2027, although the comparison is imperfect because the projects use different measure...